Abstract | ||
---|---|---|
The Speaker Recognition community that participates in NIST evaluations has concentrated on designing gender- and channel-conditioned systems. In the real word, this conditioning is not feasible. Our main purpose in this work is to propose a mixture of Probabilistic Linear Discriminant Analysis models (PLDA) as a solution for making systems independent of speaker gender. In order to show the effectiveness of the mixture model, we first experiment on 2010 NIST telephone speech (det5), where we prove that there is no loss of accuracy compared with a baseline gender-dependent model. We also test with success the mixture model on a more realistic situation where there are cross-gender trials. Furthermore, we report results on microphone speech for the det1, det2, det3 and det4 tasks to confirm the effectiveness of the mixture model. |
Year | Venue | Keywords |
---|---|---|
2011 | 12TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2011 (INTERSPEECH 2011), VOLS 1-5 | i-vectors, speaker recognition, mixture model |
Field | DocType | Citations |
I vector,Probabilistic linear discriminant analysis,Pattern recognition,Computer science,Speech recognition,NIST,Speaker recognition,Artificial intelligence,Mixture model,Microphone | Conference | 39 |
PageRank | References | Authors |
2.04 | 6 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mohammed Senoussaoui | 1 | 188 | 9.92 |
Patrick Kenny | 2 | 2700 | 214.80 |
Niko Brümmer | 3 | 595 | 44.01 |
Edward de Villiers | 4 | 135 | 8.81 |
Pierre Dumouchel | 5 | 1759 | 129.78 |